explosive welding: mixing of metals without mutual

ISSN 0031�918X, The Physics of Metals and Metallography, 2012, Vol. 113, No. 11, pp. 1041–1051. © Pleiades Publishing, Ltd., 2012.Original Russian Text © B.A. Greenberg, M.A. Ivanov, V.V. Rybin, O.A. Elkina, A.V. Inozemtsev, A.Yu. Volkova, S.V. Kuz’min, V.I. Lysak, 2012, published in Fizika Metallov iMetallovedenie, 2012, Vol. 113, No. 11, pp. 1099–1110.

1041

INTRODUCTION

Explosive welding is a very swift process and has lit�tle similarity to other methods of joining materials. Itseems outwardly simple; however, it is very complex inits physical nature; this process requires not only adetailed structural analysis, but also a new approach.With all of the variety of materials and weldingregimes, the central problem is intermixing in thetransition zone near the interface [1, 2]. Mixing occursas a result of a strong external impact, which presumeslarge plastic deformation (including pressure, shearcomponents, rotating moments of the stress, straininhomogeneity, etc.), surface friction, the cumulativejet effect, and other factors. However, it has so farremained unclear how, in such a short welding time,the metals have time to mix, even at such strong exter�nal action. This question is most urgent when consid�ering materials that have no mutual solubility.

The existence of a so�called “window of weldabil�ity” (defined in coordinates as the collision angle vs.the velocity of motion of the contact point) is the nec�essary condition for the formation of a firm joint. Toidentify the main processes that determine weldability,the structure of joints must be studied. Currently, how�ever, the role of structural studies is underestimatedand, in general, they are limited to optical microscopy.In this work, optical (OM), scanning (SEM), and

transmission electron (TEM) microscopy methods areused to study the structure of welded joints.

Previously, we investigated the welded joints of themetals (titanium–orthorhombic titanium aluminide)with normal mutual solubility in liquid and solid states[3–6]. VTI�1 and VTI�4 alloys have been selected asaluminides. Depending on the welding conditions,joints with the following various types of interfaceswere obtained:

(Aw) titanium VTI�1, wavy interface;(Ap) titanium VTI�1, flat interface, along which

melting occured;(Bw) titanium VTI�4, wavy interface;(Bp) titanium VTI�4, nearly flat, partially molten

interface.In these joints, protrusions were detected for the

first time as typical structural heterogeneities of theinterface. In some joints, melting was observed alongthe whole interface, while in others, it was observed inthe localized�melting zone with a vortex structure. Inany case, the melts present true solutions, in whichmixing occurs at the atomic level.

Next, for explosive welding, we selected copper andtantalum, which have no mutual solubility [7–9].Welding was performed at Volgograd State TechnicalUniversity. Two welding regimes were employed.Here, we present only the basic welding parameters for

Explosive Welding: Mixing of Metals without Mutual Solubility (Iron–Silver)

B. A. Greenberga, M. A. Ivanovb, V. V. Rybinc, O. A. Elkinaa, A. V. Inozemtseva, A. Yu. Volkovaa, S. V. Kuz’mind, and V. I. Lysakd

aInstitute of Metal Physics, Ural Branch, Russian Academy of Sciences, ul. S. Kovalevskoi 18, Ekaterinburg, 620990 RussiabKurdyumov Institute of Metal Physics, National Academy of Sciences of Ukraine, bul’v. Vernadskogo 36, Kiev, 03680 Ukraine

cState Polytechnical University, ul. Politekhnicheskaya 29, St. Petersburg, 195251 RussiadState Technical University, pr. Lenina 28, Volgograd, 400131 Russia

Received March 6, 2012; in final form April 4, 2012

Abstract—The results obtained for joints of dissimilar metals, iron–silver (earlier, copper–tantalum), whichform immiscible liquid suspensions, explain why they are mixed in explosive welding. Inhomogeneities of thewavy interface, such as protrusions and zones of localized melting, were observed. The effect of granulatingfragmentation, which is responsible for crushing initial materials into particles, was understood as one of themost efficient ways to dissipate the supplied energy. It is shown that, in the case of joints of metals withoutmutual solubility, zones of localized melting represent colloidal solutions, which form either emulsions orsuspensions. At solidification, the emulsion represents a hazard for joint stability due to possible separation;on the contrary, suspension can enable the dispersion strengthening of the joint. The results can be used inthe development of new metal joints without mutual solubility.

Keywords: explosive welding, mutual solubility, mixing, melting, fragmentation, colloidal solution

DOI: 10.1134/S0031918X12110075

STRUCTURE, PHASE TRANSFORMATIONS,AND DIFFUSION

1042

THE PHYSICS OF METALS AND METALLOGRAPHY Vol. 113 No. 11 2012

GREENBERG et al.

further comparison, i.e., the collision angle γ; thevelocity of motion of the contact point Vcon; the veloc�ity of collision Vcol; and the energy consumed in plasticdeformation W2, as follows:

(Cp) γ = 5.22°, Vcon = 2680 m/s, Vcol = 234 m/s,W2 = 0.33 MJ/m2; flat interface.

(Cw) γ = 11.8°, Vcon = 2125 m/s, Vcol = 440 m/s,W2 = 1.04 MJ/m2; wavy interface.

The choice of parameters (Cp) corresponds to thelower limit of the weldability range, whereas theparameters (Cw) fall into the weldability window. In[7–9], different types of heterogeneities have beenidentified, as well as typical interfaces substantial formixing in the absence of mutual solubility. Both simi�larities and differences were found in the structure oftitanium–orthorhombic titanium aluminide (referredto be below, for brevity, as titanium–aluminide). It isalso determined whether the localized�melting zones

are true solutions, as in the normal�solubility case, orif they have a different structure.

In this paper, which is a continuation of [9], thearea of the investigation of metal joints that havealmost no mutual solubility has been extended byincluding an iron–silver pair. In both cases, these arepairs of dissimilar metals that are differ greatly fromeach other with regard to their melting points andhardness. Pairs of these metals form immiscible liquidsuspensions in the molten state. We will also try toanswer why immiscible suspensions mix.

EXPERIMENTAL

Explosive welding was carried out using differentschemes and parameters at Volgograd State TechnicalUniversity. We selected the joints to be studied basedon the obtained results. In the scheme we used, theparallel arrangement of the plates was employed. Theimmobile plate was located on a metallic substrate.Welding was performed at the following parameters:

(Dw) γ = 15.6°, Vcon = 1910 m/s, Vcol = 520 m/s, W2 = 0.73 MJ/m2.

The Armco iron plate was 1.5 mm thick, while thesilver plate was 2 mm thick.

The metallographic analysis was carried out usingan Epiquant optical microscope equipped with aSIAMS computing system. The study of the micro�structure was conducted using JEM 200CX andSM�30 Super Twin transmission electron micro�scopes, Quanta 200 3D.

EXPERIMENTAL RESULTS

As mentioned above, in the previously studied cop�per–tantalum joint, both flat and wavy boundarieswere observed in different regimes of explosive weld�ing. Figure 1 shows OM and SEM images of the iron–silver interface. The wavy shape of the boundary isfairly regular (Fig. 1a), though it contains waves of dif�ferent amplitudes and even almost straight sections(Fig. 1b). The average wavelength is about 140–150 µmand the amplitude is about 50 µm. SEM images arepresented below without “SEM” indications in thefigure captions.

In the case of a wavy boundary, the interface is cor�rugated so that its longitudinal section exhibits a set ofalternating bands of iron and silver with parallelboundaries. The SEM image in Fig. 2 was taken fromthe inclined section, rather than the longitudinal sec�tion (white bands are silver). Therefore, only a fewbands are seen and protrusions are observed at theinterface. Even if the wavelength and amplitude of thewave surface change, the surface remains smooth.However, if there is a protrusion (as a cone with a sharpor smoothed apex), the surface is no longer smooth.The size of protrusions is dozens of microns; in somecases, it reaches 100 µm. As numerous SEM content

(b)

(a) 200 µm

500 µm

Ag

Fe

Fe

Ag

Fig. 1. Fe–Ag wavy interface (transverse section): (a) OMimage and (b) SEM image.


EXPLOSIVE WELDING: MIXING OF METALS 1043

measurements have shown, these protrusions are iron.Protrusions were observed for the first time in [4] in atitanium–aluminide joint, then in [7] in a copper–tantalum joint.

As can be seen from Fig. 2a, the bands can becomebroken or branched, or be divided into parts. Someareas exhibit an interface that lost its wavy character.Zones of local melting (darker than adjacent bands ofsilver) are already visible in Fig. 2a and, at a largermagnification, in Fig. 2b. These zones are clearly visi�ble in Fig. 2c, i.e., dark iron particles against a silverbackground.

Data on the chemical composition of the local�ized�melting zones (Figs. 3a, 3b) are obtained bymeans of SEM using multiple measurements. Figure 3displays the regions where measurements had beenmade. Figure 3b clearly shows iron particles againstthe silver background. The table lists the concentra�tions of elements. As can be seen from the table, thechemical composition in the center of zones isapproximately 60Ag–40Fe (at %). The structure com�position observed in the zone indicates the penetrationof elements of one material into another.

The internal structures of the localized�meltingzones are shown in Fig. 4. In Fig. 4a, the arrow marksthe region whose structure is shown in Fig. 4b at highermagnification. A similar structure can also be seen inFigs. 4c, 4d. As suggested by Figs. 4b and 4d, spheru�lites are formed during the solidification of molten sil�ver; however, ultra�rapid quenching and many nucleiwith slow growth also occur. The density of spherulitesturned out to be less in Fig. 4d, so they can be seen.One should note their quite perfect spherical shape,which minimizes the surface energy. Spherulites havedimensions of 100–200 nm. Figures 4b and 4d clearlyillustrate that the concentration of iron particles ismuch smaller (about one order of magnitude) thanthat of silver particles. At first glance, this contradictsthe results listed in the table. However, this is a spuri�ous contradiction. Thus, Fig. 3b already shows thatthe structure of the localized�melting zone is hetero�geneous; there are almost no visible particles of ironbelow square no. 5. Moreover, it follows from thesimultaneous existence of concentrated and uncon�centrated solutions that, due to ultrafast quenching,the separation of the concentrated solution may beunlikely.

Previously, among the various titanium–aluminidejoints (different compositions of aluminide and at dif�ferent regimes), spherulites were only observed in an(Ap) joint between titanium and VTI�1 alloy, and onlyif the flat interface was molten [4]. Figure 4e shows theSEM image of the structure of such an interface thatcontains spherulites. Compared to Fig. 4d, one can seethat the sizes of titanium spherulites (200–400 nm) aregreater than those of silver spherulites. As illustrated inFig. 4f, in the case of the (Cp) copper–tantalum joint,the zone of localized melting has a different structure;

(b)

(a) 500 µm

100 µm

(c) 10 µm

Fig. 2. Inhomogeneities of interface (longitudinal section):(a) protrusions; (b, c) zones of localized melting (at variousmagnifications).

1044


GREENBERG et al.

no spherulites are observed and the concentration ofboth phases are similar.

In the case of the iron–silver joint, the zone oflocalized melting presents silver containing fragmentsof iron. Here, vortices could be formed, similar tothose observed in titanium–titanium aluminide joints[3, 5]. However, no vortices were observed here, possi�bly due to the fact that the melt contained solid parti�cles, which efficiently increased its viscosity.

Figure 5 shows multiple protrusions and, as chem�ical analysis evidences, a lot of silver particles withinthe band of iron. Here, in addition, the dendrites arevisible within a band of silver, which will be consideredin detail below.

Some of the features mentioned of the interface aremore clearly visible in the relief SEM image takenafter the iron was completely etched away (Fig. 6).

This image displays silver ridges instead of bands,which are broken lines. There are also visible necksbetween the ridges, which significantly distorts theinterface. The breaks observed in the alternation ofbands in the longitudinal section, including the loss ofcontinuity, reflect the imperfection of this interface.Figure 7 shows the TEM image of a localized�meltingzone. Black voids are left by the particles of iron afterthey were etched. Judging from the observed voids,iron particles had arbitrary shapes. The electron dif�fraction pattern shows only the presence of silver.

As mentioned above, a region of silver filled withdendrites arises near the iron–silver interface, whichare known to be evidence of melting. As can be seenfrom Fig. 8a (transverse section), this region stretchesalong the entire interface, including areas of localmelting. The width of the region (vertically from theinterface) is approximately 200–250 µm. In the longi�tudinal section (Fig. 8b), there are visible single den�drites with both primary and secondary branches. Sil�ver particles are also clearly visible within the ironband, as discussed above. Figure 8b demonstrates bothdendrites and a fine�crystalline interdendritic area.Referring to Fig. 3a, note that the entire measuringregion no. 1 is composed of close packed dendriteswith some dendrites growing into the adjacent area ofsilver that avoided melting.

In the case of copper–tantalum, no dendrites wereobserved. Among the various titanium–aluminidejoints, dendrites were only observed in (Bp) titanium–VTI�4 alloy in the regime when the almost flat inter�face was partially molten. Figure 8d shows an SEM

(b)(a) 50 µm 20 µm

1

2

3

4

5

6 7

Fig. 3. Zones of localized melting (longitudinal section): (a) near interface and (b) internal zone structure; points of measure�ments are indicated.

Results of spot measurements of chemical composition ofzone of localized melting

Measuriment no. Content Ag, at % Content Fe, at %

1 97 3

2 97 3

3 0 100

4 56 44

5 62 38

6 100 0

7 2 98

Fig. 4. Zones of localized melting (longitudinal section) in (a)–(d) Fe–Ag, (e) Ti–VTI�1, and (f) Cu–Ta: (a), (c) at low magni�fication; (b, d) at high magnification; silver spherulites; (e) titanium spherulites, and (f) particles of copper and tantalum.



(b)

(c)

(a) 20 µm 1 µm

1 µm5 µm (d)

150 nm

100 nm

(e)

1 µm

1 µm

(f)

230.9 nm

114.

5 n

m

138.

5 n

m

203.

6 nm

399.

0 n

m

177. 3 n

m

1046


GREENBERG et al.

image of the structure of this region containing tita�nium dendrites.

Figure 9 exhibits (a) a cellular, (b) a band, and(c) recrystallized structures of silver that are typical ofsevere deformation. We consider that these structurescorrespond to the solid undergone no melting. Similarstructures have been observed previously in the tita�nium–aluminide [3] (Fig. 8) and copper–tantalum[7] (Fig. 10) joints.

In this paper, the microhardness was measured atvarious points in the transition zone. It was found thatthe microhardness (about 750 MPa) in the region ofthe local melting of silver containing iron particlesexceeds the minimum microhardness obtained in thesilver region with a dendritic structure (approximately450 MPa). However, this excess is far less than thatobserved previously in the case of the localized�melt�

ing zones of copper with tantalum particles (see[7], Fig. 13). In this case, the microhardness in thezone of localized melting is 4000 MPa, which is3000 MPa greater than the microhardness of copper. Itcan be assumed that stronger precipitation hardeningobserved in the case of copper–tantalum is primarilydue to the greater hardness of tantalum compared toiron and, secondarily, the higher concentration of tan�talum particles compared to iron particles.

DISCUSSION

Inhomogeneities of the Interface

Throughout this paper, we focused on the imper�fection of the wavy iron–silver interface. If onerestores the interface using the observed sections,numerous distortions in and deviations from the ideal

(b)

(a) 300 µm

100 µm

Fig. 5. Iron bands containing silver particles (longitudinalsection): (a, b) various magnifications; dendrites are visible.

500 µm

Fig. 6. Longitudinal section of the transition zone (ironetched): ridges of silver are visible.

0.2 µm

Fig. 7. TEM dark�field image in the reflection ⟨111⟩Ag ofthe zone of localized melting (iron etched): voids left byiron particles are visible.



periodic form will be found, regardless of the size ofthe protrusion, which is a key factor that determinesthe heterogeneity of the wavy interface. Due to thepresence of protrusions, regions filled with any weld�ing materials can be found on each side near the inter�face, which indicates their interpenetration. The for�mation of protrusions does not depend on the mutualsolubility of the initial metals. Protrusions wereobserved in all the joints we have examined and forboth flat and wavy interfaces. Protrusions are formedby the harder metal in the pair, i.e., aluminide, tanta�lum, or iron. A key point is the formation of protru�sions at the flat interface (Cp a copper–tantalum joint).Only the formation of protrusions can explain theobservation, in this case, of three types of regions filledwith either the original metal or a mixture thereof.

If the interface is smooth it creates problems withbonding between metals of the pair, which would

require either the reconstruction of metallic chemicalbonds, or provide the transport of point defects. How�ever, the presence of protrusions solves the problem ofsurface bonding; the protrusions play the role of nails.Another process that affects bonding is the local melt�ing of the metals near the interface. As has been shownin this and previous studies [3–9], the internal struc�ture of the localized�melting zones greatly depends onthe mutual solubility of the metals. We believe thatlocal melting is initiated by the granulating�type frag�mentation.

Granulating Fragmentation

For all joints studied, including titanium–titaniumaluminide, copper–tantalum, and iron–silver, weobtained electron microscopy images of differentzones in which there are many particles of micron and

(b)

(c)

(a) 200 µm100 µm

(d) 10 µm5 µm

Fig. 8. Areas of (a–c) silver and (d) titanium filled with dendrites: (a) transverse section (OM); (b), (c) longitudinal section (SEM)at different magnifications; (b) silver particles within the iron band; (c) fine�crystalline interdendritic region; (d) longitudinal sec�tion for (Bp) titanium–VTI�4 joint.

1048


GREENBERG et al.

nanoscaled sizes. Fragmentation into particles occursboth with and without normal solubility, no matterwhat form the interface has, i.e.,wavy or flat. In addi�tion to the particles emitted from one material toanother, continuous layers were observed that consist

of particles that interlock with each other, but insuffi�ciently tightly. This layer has been found in titanium–titanium aluminide in [5], where Fig. 7 displays theSEM image of this layer formed by titanium alu�minide. Moreover, Fig. 11 in [9] shows an SEM imageof a fragmented layer of tantalum in copper–tantalum.

Then, the question of which way do the particlesoriginate arises. Reasonable assumption is the frag�mentation of material. The possibility of a special typeof fragmentation is supposed, which we call the gran�ulating fragmentation (GF). It is a fundamentally dif�ferent type of fragmentation in comparison with thewell�known fragmentation, whose existence has beenconfirmed by numerous observations of the structureof materials upon severe deformation [10]. The classi�cal fragmentation involves piling up dislocations (andtwins), as well as the formation of coiled, cellular, andband structures and recrystallization (see Fig. 9).

The above observation of silver particles within theiron bands is the result of the GF of silver, and theobservation of iron particles in zones of localizedmelting is the result of the GF of iron. In [11], TEMobservations on micron iron particles separations fromthe interface and their movement to a distance of100 µm in a layer of aluminum are also given for thecase of iron–aluminum joints.

The consideration of the possibility of two types offragmentation is the basis for a new approach todescribing welded joint structures. Both types of frag�mentation are implemented at different distancesfrom the interface; i.e., GF occurs near the interface,where the impact is greater, and classical fragmenta�tion takes place a little further from the interface. Inour opinion, GF proceeds for a time on the order ofthe impact time, i.e., GF is a much more rapid processthan classical fragmentation. In the case of GF, thecharacteristic times are roughly estimated to be amicrosecond and, in the case of the classical fragmen�tation, they are estimated to be 108–109 µs (the struc�tural relaxation time).

Despite that the temperature in the contact areacan be quite high the internal impact is so fleeting thatthe thermally activated processes that determine themovement and rearrangement of dislocations are notpossible. It can be assumed that these processes, aswell as diffusion, can only be possible due to the effectof residual stresses and temperature. These are pro�cesses that determine the classical fragmentation,whereas the GF occurs almost without the participa�tion of dislocations. Under the rigid conditions thatare implemented in explosive welding, including thelack of relaxation mechanisms of dislocations, one ofthe most efficient channels of energy dissipation is theformation of surfaces that have the maximum totalarea. These may include free surfaces, followed bydestruction. These may be the surfaces resulting frommicrodestruction. It should be emphasized that allthese processes are athermal and can happen almostinstantaneously.

(b)

(c)

(a) 0.3 µm

0.3 µm

0.3 µm

Fig. 9. Structure of highly deformed silver: (a) cellular,(b) band, and (c) recrystallized.



GF is a process of separation into particles, whicheither fly apart or merge with each other. Possible waysof merging can be similar to those presented in [12]. Toa certain extent, GF is analogous to fragmentationunder explosion, which was studied by Mott [13]. Inboth cases, there occurs granulation and dispersion ofthe particles, but only in the case of GF, the continuityof materials is preserved because the dispersion of par�ticles takes place in a confined space. GF is a powerfulchannel for the dissipation of the supplied energy,since the surface of the particles flying apart has a largetotal area, which is the main role of the GF. As analternative process to destruction, GF increases thesurvivability of the material, which saves it fromdestruction, even under such a strong external impactas explosive welding.

Borders of the GF fragments, in our opinion, areincomplete, so that the connection between the frag�ments is weaker than between grains. This fact makesrotations and movement of the fragments possible. Itcan be expected that the processes of GF and the fric�tion between the fragments will lead to the release ofheat in abundance, which contributes to the appear�ance of melting zones near the interface.

Colloidal Solutions

The internal structure of the localized�meltingzones substantially depends on the mutual solubility ofthe constituent metals in the liquid state. As shown in[3, 5], in the case of titanium–aluminide joints, thesezones are solid solutions of titanium alloyed with nio�bium and aluminum that were originally in the alu�minide. Another situation arises if the original metals

do not have mutual solubility. In this case, these zonesof localized melting are colloidal solutions, in whichparticles of the initial elements are mixed. Here, weonly consider the case when no boiling of any of thematerials occurs in explosive welding and, thus, the gasphase does not form. In this case, one should considerthe possible forms of colloidal solutions [14]. The rela�tionship between three temperatures is significant,

i.e., is the melting point of the metal (1); is themelting point of metal (2); and Tb is the temperaturenear the interface.

Various relations are schematically shown in

Fig. 10, assuming (for simplicity) that the > .Consider the case (Fig. 10a) when

> Tb > (1)

If condition (1) is valid, then a melt of the low�meltingphase appears (hatching in Fig. 10a), which containssolid particles of the high�melting phase. This is a col�loidal solution of the suspension type. Here, theabsence of mutual solubility is of no concern and noseparation occurs. The colloidal solution is mixed dueto the circulation, which subsequently solidified into asuspension that was dispersion�strengthened by theparticles of the high�melting phase. In this case, thezones of localized melting present no hazard to thejoint stability; however, on the contrary, may contrib�ute to its strengthening.

Consider the case (Fig. 10b) when

Tb > > (2)

(1)mT (2)

mT

(1)mT (2)

mT

(1)mT (2).mT

(1)mT (2).mT

Tm(1)

Tm(2) Tm

(1)

Tm(2)

Tm(1)

Tm(2)

Tb

(a) (b) (c)

Fig. 10. Schematic representation of the various sequences of the characteristic temperatures.

1050


GREENBERG et al.

This means that both of the metals are in molten states(double hatching in Fig. 10b) within the interval Tb >

T > . The strong external impact break both liquidsinto droplets, so a colloidal solution of the emulsiontype forms. However, the emulsion is generally unsta�ble. Droplets of the same type stick together, resultingin separation into two immiscible liquids. As in thecase of so�called dilute emulsions (when the concen�tration of one phase is much smaller than the other),when the separation has no time to occur, the solidifi�

cation of one element within the range of < T < takes places and its droplets turn into solids, whereasthe second element remains liquid. In the subsequentsolidification, the colloidal solution becomes a soliddispersion�hardened suspension. In the alternativecase (Fig. 10c), when

> > Tb, (3)

neither metal melts and no colloidal solutions areformed.

Thus, in the case of colloidal solutions of immisci�ble liquids, an emulsion or suspension forms. In thesolidification, the emulsion is a danger to the continu�ity of the joint due to the possible separation; con�versely suspension can strengthen the joint.

The characteristic temperatures of the joints under

study are approximately = 3300 K, = 1300 K;

= 1800 K, and = 1200 K. If the assumptionis made that the temperature near the interface is Tb 1500–1700 K, then the relation (1) is valid for bothjoints and dispersion�strengthening suspension forms,which ensures their strength. If Tb 2000 K, then therelation (2) is valid for both joints and an emulsionforms. As can be seen from Figs. 3, 4, an emulsion isheterogeneous in various regions. The emulsion isdiluted where spherulites are observed (Fig. 4). How�ever, the emulsion shown in the image in Fig. 3 is con�centrated. These are regions where separation canappear. Here, however, no layers of iron were observed.This means that quenching is so fast that the separa�tion has no time to occur.

Dendrites

It is known that dendritic growth [15] begins whenthe liquid phase has a negative temperature gradient.In this case, this is true because the region near theinterface is the hottest. The melting occurs almostinstantaneously under explosion, but dendrites growat residual temperatures when diffusion becomes pos�sible.

The coefficient of thermal conductivity of silver isabout 418.7 W/(m K) and is one of the highest amongmetals. The thermal conductivity of iron is 74.4 W/(m K).The difference in the coefficients of thermal conduc�tivity causes the heat to sink from the interface throughthe silver. As can be seen from Fig. 8a, the thickness of

(1)mT

(2)mT (1)

mT

(1)mT (2)

mT

Ta( )mT Cu( )

mTFe( )

mT Ag( )mT

�

�

the melted layer of silver is about 200 μm. Calculationsfor determining whether there is enough energy sup�plied to melt the desired silver band are given below.For this purpose, we use the following values ofparameters for silver: an atomic weight of 107.9, den�sity of 10.5 g/cm3, melting temperature Tm = 1234 K,melting heat Qm = 11.34 kJ/mol, and heat capacity

≈ 24 J/(mol K).Hence, it is easy to show that the energy required to

melt this layer Em is approximately 0.23 MJ/m2, andthe energy required to heat it Eh to 1000 K is approxi�mately 0.48 MJ/m2. Thus, the total energy required toheat and melt a 200�μm�thick silver layer is Et ≈0.71 MJ/m2.

This estimate is too high, since the total energyinjected during the explosion of the iron–silver systemis only 0.73 MJ/m2. However, here, the energyrequired to fragment and heat iron and silver awayfrom the melting zone, etc. was not considered. How�ever, it has also not been taken into account that thedendrite layer has a thickness of less than 200 μm inmany places, which may reduce the energy requiredfor melting. Thus, in general, we can assume that theenergy balance is sufficient for heating and melting thesilver layer.

CONCLUSIONS

(1) A structural study of the joints of metals withoutmutual solubility (copper–tantalum and iron–silver)resulted in a foundation in which, under explosivewelding, the interpenetration of the materials is car�ried out by the formation of protrusions, the ejectionof particles of one material into another, and the for�mation of zones of localized melting.

(2) Granulating fragmentation is observed. GF isthe process of separation into particles, which eitherfly apart or merge with each other. The role of GF inexplosive welding was understood to be one of themost efficient ways of dissipating the energy supplied.To a certain extent, GF is analogous to the fragmenta�tion at the explosion, which was studied by Mott. GFresults from microdestructions and is an alternativeprocess to destruction.

(3) It was shown that, in the case of joints of metalswith normal solubility, the localized�melting zonesformed true solutions, whereas in the case of metaljoints without mutual solubility, the localized�meltingzones formed colloidal solutions.

(4) It was shown that the investigated joints couldform either emulsions or dispersion�strengthened sus�pensions. When solidified, the emulsion presents ahazard to joint stability due to possible separation;conversely, suspension can contribute to the strength�ening the joint. Thus, the risk areas that can arise inthe absence of mutual solubility were identified. Theresults can be used to develop new metal joints withoutmutual solubility.

Cv



ACKNOWLEDGMENTS

Electron microscopic studies were performed atthe Electron Microscopy Center of CollaborativeAccess, Urals Branch, Russian Academy of Sciences.This work was supported in part by the Russian Foun�dation for Basic Research, project no. 10�02�00354;by projects 12�2�006�UT, 12�U�2�1011, 12�2�2�007of the Ural Branch of the Russian Academy of Sci�ences; and by the National Target Program of Ukraine“Nanotechnologies and Nanomaterials,” no. 1.1.1.3�4/10�D.

REFERENCES

1. A. A. Deribas, Physics of Strengthening and Welding byExplosion (Nauka, Novosibirsk, (1980) [in Russian].

2. V. I. Lysak and S. V. Kuzmin, Explosion Welding (Mash�inostroenie, Moscow, 2005) [in Russian].

3. V. V. Rybin, I. I. Sidorov, B. A. Greenberg,O. V. Antonova, O. A. Elkina, A. V. Inozemtsev,A. M. Patselov, and M. A. Ivanov, “Formation of Vorti�ces during Explosive Welding (Titanium–Orthorhom�bic Titanium Aluminide),” Phys. Met. Metallogr. 108,353–364 (2009).

4. V. V. Rybin, B. A. Greenberg, M. A. Ivanov,S. V. Kuz’min, V. I. Lysak, O. A. Elkina, A. M. Patselov,A. V. Inozemtsev, O. V. Antonova, and V. E. Kozhevni�kov, “Structure of the Welding Zone between Titaniumand Orthorhombic Titanium Aluminide for ExplosionWelding: I. Interface,” Russ. Metall. (Metally) 2011(10), 1008–1015 (2011).

5. B. A. Greenberg, M. A. Ivanov, V. V. Rybin,S. V. Kuz’min, V. I. Lysak, O. A. Elkina, A. M. Pat�selov,, O. V. Antonova, and A. V. Inozemtsev, “Struc�ture of the Welding Zone between Titanium andOrthorhombic Titanium Aluminide for ExplosiveWelding: II. Local Melting Zones,” Russ. Metall.(Metally) 2011 (10), 1016–1025 (2011).

6. V. V. Rybin, B. A. Greenberg, M. A. Ivanov, A. M. Pat�selov, O. B. Antonova, O. A. Elkina, A. V. Inozemtsev,and G. A. Salishchev, “Nanostructure of Vortex duringExplosive Welding,” J. Nanosci. Nanotechnol. 11,8885–8895 (2011).

7. B. A. Greenberg, M. A. Ivanov, V. V. Rybin,O. A. Elkina, A. M. Patselov, O. B. Antonova,A. V. Inozemtsev, G. A. Salishchev, and V. E. Kozhevni�kov, “Transition Zone Structure and Its Influence onthe Strength of Copper–Tantalum Explosion WeldedJoint,” Deform. Razrush. Mater., No. 9, 34–40 (2011).

8. B. A. Greenberg, O. A. Elkina, O. V. Antonova,A. V. Inozemtsev, M. A. Ivanov, V. V. Rybin, andV. E. Kozhevnikov, “Peculiarities of Formation ofStructure in the Transition Zone of Cu–Ta Joint Madeby Explosion Welding,” The Paton Welding J., No. 7,20–25 (2011).

9. B. A. Greenberg, M. A. Ivanov, V. V. Rybin,A. V. Inozemtsev, O. V. Antonova, O. A. Elkina,A. M. Patselov, S. V. Kuzmin, V. I. Lysak, andV. E. Kozhevnikov, “Inhomogeneities of the InterfaceProduced by Explosive Welding,” Phys. Met. Metal�logr. 113, 176–189 (2012).

10. V. V. Rybin, Large Plastic Deformations and Fracture ofMaterials (Metallurgiya, Moscow, 1986) [in Russian].

11. L. I. Markashova, V. V. Arsenyuk, G. M. Grigorenko,and E. N. Berdnikova, “Peculiarities of Mass TransferProcesses at Different Metal Welding by Pressure,”Svaroch. Proizvod., No. 4, 28–35 (2004).

12. R. Z. Valiev and I. V. Aleksandrov, Bulk NanostructuralMetal Materials (Akademkniga, Moscow, 2007)[in Russian].

13. N. F. Mott, “Fragmentation of Shell Cases,” Proc. Roy.Soc. A 189, 300–308 (1947).

14. B. D. Summ, Fundamentals of Colloid Chemistry (Aka�demiya, Moscow, 2009) [in Russian].

15. M. Flemings, Solidification Processing, (New York:McGraw�Hill, 1974).

explosive welding: mixing of metals without mutual

Documents